Parallel algorithm of navier-stokes model for magnetic nanoparticles drug delivery system on distributed parallel computing system
Integrated mathematical Navier-Stokes model for transportation of drug across the blood flow medium by partial differential equations (PDE) with one dimensional (1D) and two dimensional (2D) parabolic type in cylindrical coordinates system are considered. The process of magnetic nanoparticle drug de...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2017
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/77901/1/SakinahAbdulHananMFS20171.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Integrated mathematical Navier-Stokes model for transportation of drug across the blood flow medium by partial differential equations (PDE) with one dimensional (1D) and two dimensional (2D) parabolic type in cylindrical coordinates system are considered. The process of magnetic nanoparticle drug delivery system is made measurable by identifying some parameter such as magnetic nanoparticle targeted delivery, blood flow, momentum transport, density and viscosity on drug release through blood medium, the intensity of magnetic fields, the radius of the capillary and controllability expression to control the concentration of blood. Finite difference method (FDM) with centre difference formula was used to discretization the mathematical model. This research focuses on two types of discretization controlled by weighted parameter 6 = 1 and 6 = - which are implicit (IMP) and Crank Nicolson (CN) schemes respectively. The implementation of several numerical iterative methods such as Alternating Group Explicit (AGE), Red Black Gauss Seidel (RBGS) and Jacobi (JB) method are used to solve the linear system equation (LSE) and is one of the contributions of this research. The sequential algorithm was developed by using C Microsoft Visual Studio 2010 Software. The numerical result was analysed based on execution time, number of iteration, maximum error, root mean square error, and computational complexity. The grid generation process involved fine grained of large sparse matrix by minimizing the size of interval, increasing the dimension of model and level of time steps. Parallel algorithm was proposed for increasing the speedup of computations and reducing computational complexity problem. The parallel algorithms for solution of large sparse systems were design and implemented supported by the distributed parallel computing system (DPCS) containing 8 processors Intel CORE i3 CPUs employing the Parallel Virtual Machine (PVM) software. The parallel performance evaluation (PPE) in term of execution time, speedup, efficiency, effectiveness, temporal performance, granularity, computational complexity and communication cost were analysed for the performance of parallel algorithm. As a conclusion, the thesis proved that the 1D and 2D Navier-Stokes model is able to be parallelized and parallel AGE method is the alternative solution for the large sparse simulation. Based on numerical result and PPE, the parallel algorithm is able to reduce the execution time and computational complexity compared to the sequential algorithm. |
|---|